Learning convex inference of marginals

Justin Domke
Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence, PMLR R6:137-144, 2008.

Abstract

Graphical models trained using maximum likelihood are a common tool for probabilistic inference of marginal distributions. However, this approach suffers difficulties when either the inference process or the model is approximate. In this paper, the inference process is first defined to be the minimization of a convex function, inspired by free energy approximations. Learning is then done directly in terms of the performance of the inference process at univariate marginal prediction. The main novelty is that this is a direct minimization of empirical risk, where the risk measures the accuracy of predicted marginals.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR6-domke08a, title = {Learning convex inference of marginals}, author = {Domke, Justin}, booktitle = {Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence}, pages = {137--144}, year = {2008}, editor = {McAllester, David A. and Myllymäki, Petri}, volume = {R6}, series = {Proceedings of Machine Learning Research}, month = {09--12 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r6/main/assets/domke08a/domke08a.pdf}, url = {https://proceedings.mlr.press/r6/domke08a.html}, abstract = {Graphical models trained using maximum likelihood are a common tool for probabilistic inference of marginal distributions. However, this approach suffers difficulties when either the inference process or the model is approximate. In this paper, the inference process is first defined to be the minimization of a convex function, inspired by free energy approximations. Learning is then done directly in terms of the performance of the inference process at univariate marginal prediction. The main novelty is that this is a direct minimization of empirical risk, where the risk measures the accuracy of predicted marginals.}, note = {Reissued by PMLR on 09 October 2024.} }
Endnote
%0 Conference Paper %T Learning convex inference of marginals %A Justin Domke %B Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2008 %E David A. McAllester %E Petri Myllymäki %F pmlr-vR6-domke08a %I PMLR %P 137--144 %U https://proceedings.mlr.press/r6/domke08a.html %V R6 %X Graphical models trained using maximum likelihood are a common tool for probabilistic inference of marginal distributions. However, this approach suffers difficulties when either the inference process or the model is approximate. In this paper, the inference process is first defined to be the minimization of a convex function, inspired by free energy approximations. Learning is then done directly in terms of the performance of the inference process at univariate marginal prediction. The main novelty is that this is a direct minimization of empirical risk, where the risk measures the accuracy of predicted marginals. %Z Reissued by PMLR on 09 October 2024.
APA
Domke, J.. (2008). Learning convex inference of marginals. Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R6:137-144 Available from https://proceedings.mlr.press/r6/domke08a.html. Reissued by PMLR on 09 October 2024.

Related Material